New Relic, the Intelligent Observability company, announced two complementary innovations, Agentic AI Monitoring and the New Relic AI Model Context Protocol (MCP) Server, that together transform system complexity into clarity and help businesses deliver more perfect software in the AI age.
Agentic AI Monitoring provides businesses with holistic visibility into interconnected agents and tools so they can optimise their agentic workforces. The New Relic AI MCP Server opens up a standardised way of enabling a powerful ecosystem of agents that fuel advanced agentic workflows. This integration allows popular AI assistants like GitHub Copilot, ChatGPT, Claude, and Cursor to access detailed New Relic observability data directly, embedding critical insights into engineers’ workflows.
“The convergence of AI workloads, cloud-native architectures, and real-time data processing has created a perfect storm of complexity,” said New Relic Chief Product Officer Brian Emerson. “Our platform uses intelligent automation and unified data correlation to diffuse that complexity so you can operate your business confidently and at scale. Our latest innovations further empower enterprises to adopt AI systems that create real business value, rather than cutting into the bottom line.”
Without modern observability built for the AI era, silent issues can ripple through LLM-powered applications and agentic AI systems unnoticed. The 2025 Observability Forecast found the majority of organizations are aware of the challenges — the use of AI monitoring capabilities went from 42% in 2024 to 54% in 2025 — as the cost of digital business downtime grows. The research shows high-impact outages now carry a median cost of $2 million USD per hour, or more than $33,000 for every minute systems remain down.
Observability for Agentic AI
As organisations adopt sophisticated agentic AI, complexity grows quickly: agents rely on each other’s outputs from multiple MCP servers, and shared context and memory, creating intricate webs that are difficult to debug. If one agent hallucinates and another is pulling its information, the issue might not show up for several steps. To unlock value from agentic workforces and allow teams to build with confidence, teams need full visibility into these agent-level interactions and new ways to evaluate agents’ work. This will allow them to quickly identify and resolve issues, and ensure ROI on their organisation’s investment.
New Relic Agentic AI Monitoring provides visibility into every agent and tool call within multi-agent collaborations. Organisations now understand how their mesh of agents communicate with each other, so they can troubleshoot faster and avoid downtime. The capability delivers granular insights into tool utilisation, performance, and errors with a view that shows which agents and tools were called — in what order — and key performance data. Users gain a consolidated AI Inventory view of agents and tools as well as detailed insights into their names, latency, and errors. An Agents Service Map visualises inter-agent interactions, and offers the ability to drill down into individual agent performance details and their traces.
Unlike solutions limited to monitoring individual LLMs, New Relic provides holistic observability across interconnected agents and tools and, critically, also the services and infrastructure they rely on. This enables engineering and DevOps teams to pinpoint issues faster, accelerate root cause analysis, and optimise performance across their entire AI-enabled stack. The capability is only possible by building Agentic AI Monitoring on top of tried-and-true APM and infrastructure monitoring tools like New Relic’s.
New Relic AI MCP Server
The New Relic AI MCP Server brings New Relic’s capabilities to AI agents, unlocking a crucial ecosystem of integrated tools. Many powerful AI assistants currently operate without key insights into the performance of code running in production. This limits their usefulness and forces developers to toggle between platforms. With the New Relic AI MCP server, engineers can retrieve deep, detailed data and insights from wherever they’re working, making their favorite AI agents instantly more powerful and productive. By eliminating context switching, teams can respond to incidents faster, shorten MTTR, improve uptime, and accelerate time-to-market. Every engineer can now ask questions and receive actionable insights promptly, democratising observability across the organisation.
“As enterprises deploy agentic AI to accelerate software delivery, engineers have lacked direct access to observability data within their workflows,” said IDC Group Vice President Stephen Elliot. “Observability platforms will need to fill that gap by making observability capabilities available to any MCP-compatible agent. By creating an intelligent feedback loop where AI systems become more observable and reliable, while observability platforms become more intelligent and proactive, the industry can equip enterprises with the confidence needed to innovate at speed.”
Outlier Detection
New Relic is also introducing Outlier Detection, which works in tandem with anomaly detection to detect and analyse aberrant behaviors. New Relic Outlier Detection highlights data points that signal issues or failures, enabling teams to prioritise workstreams and address incidents proactively before they impact end-users. The solution goes beyond industry-standard algorithms to not only flag outliers but also streamline remediation.
Availability
Limited previews of Agentic AI Monitoring, the New Relic AI MCP Server, and Outlier Detection are now available as part of the New Relic Intelligent Observability Platform.
